This project is designed to understand the rapid increase in autism prevalence in the United States over the past two decades. Proposed are new analyses of complex, multi- level temporally sensitive data sets that will enable me to determine the extent to which familial, environmental, gene-environment, and diagnostic drift/substitution are driving the autism epidemic. Detailed attention to and models capable of capturing social network and social influence underpinnings of the epidemic are developed. New analysis models for intercalating spatial and social network data are developed. These models where appropriate are extended to a wide range of developmental disorders that have increased rapidly in prevalence.